Health services research demonstrating unexplained variation in use of more expensive health services has been a primary motivation for studying the comparative effectiveness of alternative treatments (1, 2). For example, the Congressional Budget Office report that argued for a greater federal role in comparative effectiveness research cited markedly greater geographic variation in hospitalization for back surgery (a more elective procedure) than in hospitalization for hip fracture (a non-elective procedure) (2). The limited practice variation research in mental health has also demonstrated significant variation between geographic areas or provider groups in rates of hospital care, outpatient services, and use of specific medications (3-6). Some ofthis variation has been associated with differences in patients'demographic characteristics and in supply of providers or facilifies. These limited data suggest that practice variation in mental health care is as much a concern as is variation in use of medical or surgical treatments. We identify two issues not adequately addressed by previous research regarding health care practice variation. First, previous studies have lacked the multi-level data necessary to disentangle how patient, provider, and health system characteristics are related to variation in treatment choices and subsequent health services costs. Differences between outpatient clinics in adoption of a new medication might represent differences in the patients served, differences in providers'preferences or pracfice styles, or health system factors (such as formulary policies or medication management programs). Second, identifying variation in service use does not determine whether higher or lower rates are more desirable. As discussed below, identifying providers or facilities with higher rates for a particular treatment has very different implications depending on whether that treatment improves health outcomes or simply increases costs. Variation in care is of interest in terms of both quality of care and restraining the growth of health care costs. In the case of health services demonstrated to improve health outcomes (high-value services), significant variation in care suggests potential under-ufilizafion of effective treatments. In such cases, we would hope that variations in treatment choices would be explained either by differences in clinical need (severity of illness, prior treatment experience, co-occurring conditions, etc) or by differences in patients'preferences. Ideally, use of effective treatments would be only minimally infiuenced by non-clinical factors such as providers'preferences or patients'insurance coverage. For example, variation between racial or ethnic groups in receipt of effective or high-value services might be appropriate if that variation refiected racial or ethnic differences in patients'treatment preferences. But racial or ethnic variation in receipt of high-value treatments would not be appropriate if it instead refiected providers'biases. In the case of health services that increase costs without demonstrated clinical benefit (low-value services) significant variation in care suggests potential over-use of services and an opportunity to decrease costs without sacrificing health outcomes. We might hope that use of low-value services is more infiuenced by cost-reduction efforts (differential copayments, formulary restrictions, organizational initiatives). And if we observe significant variation among providers in use of low-value services, this suggests an opportunity for additional education or quality improvement efforts to increase the efficiency of care. Eariy adoption of a new treatment has quite different implications for quality improvement and policy depending on whether that treatment improves outcomes or simply increases costs. Because previous research on practice variation has not typically distinguished between these two scenarios, we know little about whether rates of use for high- and low-value treatments tend to vary together between providers, facilities, or geographic areas - or whether high- and low-value services show distinct patterns.
| Coleman, Karen J; Johnson, Eric; Ahmedani, Brian K et al. (2018) Predicting Suicide Attempts for Racial and Ethnic Groups of Patients During Routine Clinical Care. Suicide Life Threat Behav : |
| Lu, Christine Y; Penfold, Robert B; Toh, Sengwee et al. (2018) Near Real-time Surveillance for Consequences of Health Policies Using Sequential Analysis. Med Care 56:365-372 |
| Simon, Gregory E; Johnson, Eric; Lawrence, Jean M et al. (2018) Predicting Suicide Attempts and Suicide Deaths Following Outpatient Visits Using Electronic Health Records. Am J Psychiatry 175:951-960 |
| Rossom, Rebecca C; Simon, Greg E; Coleman, Karen J et al. (2018) Are wishes for death or suicidal ideation symptoms of depression in older adults? Aging Ment Health :1-7 |
| Lu, Christine Y; Simon, Gregory; Soumerai, Stephen B et al. (2018) Counter-Point: Early Warning Systems Are Imperfect, but Essential. Med Care 56:382-383 |
| Lu, Christine Y; Simon, Gregory; Soumerai, Stephen B (2018) Counter-Point: Staying Honest When Policy Changes Backfire. Med Care 56:384-390 |
| Waitzfelder, Beth; Stewart, Christine; Coleman, Karen J et al. (2018) Treatment Initiation for New Episodes of Depression in Primary Care Settings. J Gen Intern Med 33:1283-1291 |
| Becerra, Tracy A; Massolo, Maria L; Yau, Vincent M et al. (2017) A Survey of Parents with Children on the Autism Spectrum: Experience with Services and Treatments. Perm J 21: |
| Zeber, John E; Coleman, Karen J; Fischer, Heidi et al. (2017) The impact of race and ethnicity on rates of return to psychotherapy for depression. Depress Anxiety 34:1157-1163 |
| Madden, Jeanne M; Lakoma, Matthew D; Lynch, Frances L et al. (2017) Psychotropic Medication Use among Insured Children with Autism Spectrum Disorder. J Autism Dev Disord 47:144-154 |
Showing the most recent 10 out of 49 publications
